New Research: 5 Federal AI Challenges Workday Solves For

AI has moved from a Washington talking point to a foundational part of federal agency mission achievement. In a new survey of defense and civilian agency leaders by Workday Government and GovExec, 42% reported that AI is having a positive impact on their work—with gains in data analysis and reporting (79%), mission delivery (69%), and procurement and finance (67%).

At the same time, only 13% said they are using agentic AI broadly across departments. That means the biggest gains are still on the table.

What's preventing agencies from embedding AI more deeply into their operations? The survey reveals five of the biggest barriers—and how Workday Government Cloud helps agencies overcome them to achieve widespread adoption and mission success.

1. Security Vulnerabilities 

According to the survey, security is federal leaders’ top concern about agentic AI. Eighty-five percent said they are very or extremely concerned about vulnerabilities introduced by agents—understandbale for agencies that operate under zero-margin-for-error accuracy requirements against a threat landscape growing more sophisticated by the month. 

Yet the majority of those same leaders—76%—also saw gains in cybersecurity and IT operations from AI. The difference comes down to software with security built in, not bolted on. 

Traditional AI implementations often force agencies to bolt on third-party apps, creating new attack surfaces and compliance exposure. Workday Government Cloud takes a different approach. AI and machine learning (ML) are built into the platform's unified core architecture, and the entire Workday Government Cloud is FedRAMP-authorized—with zero-trust design, AES 256-bit encryption in transit and at rest, and always-on audit trails. Agents access only what their assigned identity permits, so human users and AI operate under the same guardrails.

2. Bias and Transparency

Bias and discrimination ranked second among federal leaders' agentic AI concerns. Every human capital manager surveyed—100%—said they are at least moderately concerned about biased AI outputs, and 56% cited lack of transparency and explainability as the reason why. The stakes are high: if agencies can't see how a model reached a recommendation, they can't defend it to citizens, oversight bodies, or their own workforce.

Building transparency and monitoring for bias starts with clear documentation of what data models agents are allowed to use and how. Just as important, agencies must retain control over their own data and govern AI agents with the same accountability applied to human employees.

For example, every Workday Government Cloud AI feature ships with clear documentation of what data it uses and who is allowed access. Agencies control whether to contribute data for ML training, customer-provided training data is purged within 30 days, and Workday never shares customer data with other vendors to train public models. Granular, feature-level opt-in and opt-out controls give agencies the visibility they need to trust and defend every AI-assisted decision. 

3. Workforce Anxiety 

Just a year ago, headlines predicted AI would eliminate large swaths of the federal workforce. The research tells a different story: only 13% of federal leaders reported headcount reductions tied to AI, and just 7% expect role eliminations. Meanwhile, 27% said employees now spend more time on higher-value work. 

Successful AI adoption is as much about people as it is about technology. Agency leaders should build internal programs that help employees adapt to new workflows and train them on new AI tools, so agents and workers can collaborate effectively. In practice, that often means   agents handle  repetitive work while  employees bring mission context and judgement.

To support that collaboration, Workday Government Cloud employs a strict ‘human-in-the-loop’ AI architecture that keeps people in control of critical administrative functions. AI cannot  execute actions autonomously without explicit human authorization—federal employees review and edit machine-generated recommendations before  changes take effect. Comprehensive audit trails track user interactions, ensuring full accountability and transparency across all automated workflows. And a centralized Agent System of Record serves as a single source of truth to govern, manage, and optimize an organization’s entire fleet of native and third-party AI agents alongside its human workforce. 

This human-centric governance model mitigates risk by putting human judgment where it matters most: identifying algorithmic bias before it affects decisions. By keeping the final decision-making authority with agency personnel, this approach fosters organizational trust in AI while leveraging continuous human feedback to improve its accuracy over time.

4. AI Policy Oversight

Only 22% of federal leaders in our survey listed internal AI governance policies as their primary bias-mitigation method. Just 32% require formal human review of AI outputs before action is taken. And in too many agencies, oversight policies are still being developed even as agents move into production.

Closing this gap means treating governance as a foundational first step—putting policies in place before agents scale, not after. 

That becomes far easier when governance is already built into the software agencies use. For example, Workday Government Cloud helps agencies operationalize governance rather than just document it. Administrators can toggle individual AI features on or off and review a full ML inventory through the data contribution configuration interface. And because Workday’s Agent System of Record gives agencies visibility into all their agents across the business, leaders can enforce policies consistently across HR, finance, and payroll—eliminating the shadow-AI risk that comes with disconnected point solutions.

5. Bad Data and Lawless AI

Over half of the federal organizations we surveyed said they are still early in their agentic AI journey. Many want to move faster but don’t trust the underlying data. Fragmented legacy systems produce inconsistent and overlapping data—and layering AI on top of bad data only compounds the problem. This is what fuels lawless AI: systems that produce unreliable, misleading, or unintended outcomes because the data they access is incomplete or inconsistent. As federal AI adoption accelerates, scaling agents on top of poor-quality data means scaling flawed, untrustworthy systems.

Workday Government Cloud solves these critical data issues by managing data within a single, unified platform. This unified approach ensures that data is governed consistently across essential functions like HR, finance, and payroll so AI interacts with the same data,  the same way, every time. That standardization reduces errors and establishes repeatable processes, making AI outcomes easier to verify, audit, and trust. A clean data foundation also enables agencies to safely automate routine, repetitive tasks, freeing up government employees for complex work, while an always-on audit capability tracks what actions were taken, when, why, and under what rules.

From Pilots to Mission Impact

Federal agencies are all in on AI—just 4% report no AI in their operations at all. But the focus is shifting. If 2026 was the year of pilots, 2027 will be the year agencies move beyond experimentation and turn early wins into lasting mission impact. And Workday Government Cloud is built to help them get there. 

Methodology

Market Connections deployed a 9-question poll to a random sample of 200 US federal civilian and defense IT employees. The poll was fielded in June 2026. 

About Workday Government

Workday Government is the enterprise AI platform for managing people, finance, and agents. Workday Government unifies HR and finance on one intelligent platform with AI at the core to empower public sector organizations at every level with the clarity, confidence, and insights they need to adapt quickly, make better decisions, and deliver on their missions. For more information, including open roles where you can support our Federal agency customers, visit workday.com/federal.

About Market Connections

As GovExec’s research division, Market Connections is dedicated to advancing the business of government through analysis, insight, and analytical independence. An extension of Government Executive’s 50 years of exemplary editorial standards and commitment to the highest ethical values, Market Connections studies influential decision makers from across government to produce intelligence-based research and analysis.

Download the full report to see how  federal defense and civilian leaders are moving agentic AI from experimentation to execution.